Improving GNSS data accuracy: Implementing Non-tidal Corrections at the Observation-level in GNSS data

Non-tidal loading (NTL) displacements contribute to stochastic variability in Global Navigation Satellite Systems (GNSS) coordinate time series. In this study, we investigate the influence of non-tidal atmospheric, atmospheric-ocean, and atmospheric-ocean-hydrology loading displacement corrections on the stochastic characteristics of GNSS 3D time series at permanent GNSS stations in Finland. We apply corrections for NTL at the observation-level as well as post-processing. We qualitatively compared the time-series statistics of NTL-corrections and observed weighted RMS reduction rates of up to 20% and 58% for the horizontal and vertical coordinates, respectively. After applying NTL corrections, substantial improvements were evident in the vertical component, with the power-law noise amplitude nearly halved from approximately $$\:10\:mm/y{r}^{-\kappa\:/4}$$ to below $$\:5\:mm/y{r}^{-\kappa\:/4}$$ , while the white noise amplitude remained zero for observation-level (consistent with the original residual white noise amplitude). Trend uncertainties in the vertical component were also reduced by nearly half. Additionally, our results demonstrate for first the time that integrating hydrological loading products into the corrections led to annual phase changes, enhancing the overall accuracy of GNSS time series. The findings highlight the critical role of NTL-corrections in improving the precision of GNSS measurements, thereby contributing to more accurate monitoring and understanding of geophysical processes in Northern Europe.